Speech, Language and Hearing

ISSN: 2050-571X (Print) 2050-5728 (Online) Journal homepage: http://www.tandfonline.com/loi/yslh20

The bilingual voice: Vocal characteristics when

speaking two languages across speech tasks

Binna Lee & Diana Van Lancker Sidtis

To cite this article: Binna Lee & Diana Van Lancker Sidtis (2017): The bilingual voice: Vocal
characteristics when speaking two languages across speech tasks, Speech, Language and
Hearing, DOI: 10.1080/2050571X.2016.1273572

To link to this article: http://dx.doi.org/10.1080/2050571X.2016.1273572

Published online: 03 Jan 2017.

Submit your article to this journal

Article views: 28

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=yslh20

Download by: [Inst for Basic Research] Date: 13 January 2017, At: 10:29
The bilingual voice: Vocal characteristics
when speaking two languages across speech
tasks
Binna Lee 1,2 , Diana Van Lancker Sidtis 1,2
1
Department of Communicative Sciences & Disorders, New York University, New York, NY, USA, 2Brain and
Behavior Laboratory, Geriatrics Division, Nathan Kline Institute for Psychiatric Research, Orangeburg, NY, USA

The present study compares vocal parameters in bilinguals’ use of two languages across different speech
tasks, in order to examine whether language selection and task type will manifest different outcomes.
Korean-English (KE) and Mandarin-English (ME) speakers performed three speech tasks – a reading
passage, a monologue, and a picture description – in their two spoken languages. Fundamental
frequency (F0), F0 variability, intensity, and speaking rate were measured. Results yielded an effect of
language on average F0s, as both bilingual groups showed a significantly higher F0 in Mandarin or
Korean compared to English. For the KE group, a higher F0 appeared in Korean compared to English
across all tasks, while for the ME group, a difference in F0 between Mandarin and English was found for
reading only. Both groups had the highest F0s in the reading task. There was a difference in F0 variability
between two spoken languages only for the KE group monologue. Mean intensity of the ME group was
overall higher than that of the KE group. Speaking rate was significantly slower in English than the native
language for both groups, but the KE group demonstrated a greater difference in pace between their two
spoken languages. Findings demonstrated that bilingual speakers produced notably different voice
patterns contingent on language and speech task, indicating that inter- and intra-speaker variability in
speakers’ vocal features can be attributed in part to language effects.
Keywords: Voice, Task effect, Fundamental frequency, Bilingualism, Korean, Mandarin

Introduction speech. The voice pattern broadly considered mani-

According to the US census Bureau and 2011
American Community Survey, more than 350 different
languages are spoken in US homes. In great metropo-
litan areas such as New York or Los Angeles, approxi-
mately 38 –54% of residents are reported to speak a
language other than English at home. These facts
reveal increasing diversity of language groups in the
US followed by enhanced awareness of the impact of
these spoken languages on aspects of individuals’ per-
formances. The current study expands this line of
interest on diversity of spoken languages through an
investigation of voice profiles of bilingual speakers
and examines the vocal influence of language selection
and speech task on their speech performance.
The term ‘voice’ has many definitions, depending on
context and perspective. Kreiman and Sidtis (2011)
distinguish between voice narrowly considered, refer-
encing laryngeal-supraglottal phonation, and voice
broadly considered, which is synonymous with
Correspondence to: Binna Lee, Steinhardt, Department of Communicative
Sciences and Disorders, New York University, 665 Broadway, 9th Floor,
New York, NY 10012 USA.
Email: bnlee@nyu.edu
fests auditory-acoustic parameters such as fundamen-
tal frequency mean and variation, intensity mean
and variation, temporal cues including pausing and
rate, voice quality, such as harsh or breathy, and
articulatory features of pronunciation, all of which lis-
teners use to make judgments about the speaker’s per-
sonal characteristics and intentions. In this study,
‘voice’ refers the vocal pattern broadly considered,
with a focus on fundamental frequency (F0), intensity,
and rate. The other variable in this study, the bilingual,
can also benefit from a definition. The term bilingual-
ism is defined here as the ability to use two different
languages with a range of linguistic abilities (speaking,
reading, and writing) in different social contexts
(Roberts and Shanker, 2007).
Over the past few decades, a number of voice studies
have undertaken empirical investigations to explore
the effect of language on the acoustic aspects of
speech focusing on a variety of different language
groups (Andrianopoulos et al., 2001; Altenberg and
Ferrand, 2006; Awan and Mueller, 1996; Hudson
and Holbrook, 1982; Jarvinen et al., 2013; Mayo

© 2017 Informa UK Limited, trading as Taylor & Francis Group

DOI 10.1080/2050571X.2016.1273572 Speech, Language and Hearing 2017 1
Lee and Sidtis The bilingual voice
and Manning, 1994; Mennen et al., 2012; Morris,
1997; Ng et al., 2012; Sapienza 1997; van Bezooijen,
1995; Xue et al., 2001; Wheat and Hudson, 1988). In
a majority of these comparative studies, F0 differences
have been the focus of interest. In one of the earlier
studies, for example, Hanley et al. (1966) compared
acoustic values of F0 in speakers of Tagalog,
Spanish, Japanese and English. The results indicated
that the English group had lower F0 measures com-
pared with Japanese and Spanish groups. In another
study, Andrianopoulos et al. (2001) compared F0s in
isolated vowel productions in speakers of English,
Mandarin, and Hindi, and found that Mandarin
speakers had significantly higher F0 values compared
to the other speakers. Japanese and Dutch women also
exhibited different acoustic speech patterns, wherein
the former speakers used a higher range of F0 (van
Bezooijen, 1995). Similar findings of F0 differences
among other ethnic groups were observed; elderly
Euro-American females had a significantly a higher
mean F0 than African-American females (Xue et al.,
2001). More recently, Keating and Kuo (2012) com-
pared the F0 of Mandarin and English monolinguals
and found that Mandarin speakers had a higher F0
and larger F0 range than English speakers. As
shown, different language groups exhibited difference
in F0 parameters.
However, other studies failed to find difference in
F0s across various language and ethnic groups. For
example, Sapienza (1997) reported no significant
differences in mean F0s in African-American and
Caucasian speakers when producing English vowels.
Awan and Mueller (1996) further compared mean
F0 measures in connected speech among African-
American, Caucasian -American, and Hispanic
kindergartners. In their findings, except for African-
American children who had a significantly lower F0
measures than Hispanic children, the remaining
groups revealed comparable F0s. Not only average
F0s, but F0 variability revealed no differences
among different language groups (Baken and
Orlikoff, 2000; Gelfer and Denor, 2014).
Although these findings are inconsistent, some
studies do reveal a relationship between language
and vocal attributes. There has been a heavy focus
on F0, and not all contributory vocal characteristics
have been examined. It is noteworthy that intensity,
to our knowledge, has not been compared across
languages, although loudness is a strong communica-
tor of personal states and traits, such as anger and
aggression (Kreiman and Sidtis, 2011). It is likely
that linguistic groups may differ on this parameter.
Another strongly signaling vocal cue, rate, has received
only a little attention in this domain of study, even
though syllable rate differences between language
2 Speech, Language and Hearing 2017
types are impressionistically quite obvious (Amino
and Osanai, 2015).
The studies reviewed above compared voices of
monolingual speakers, so that the observed differences
in vocal measures might be attributable to anatomical
factors rather than to cross-linguistic differences (Ng
et al., 2012; Todaka, 1993). To exclude the possibility
of anatomical or morphological differences from eth-
nicities as a confounding factor, recent studies exam-
ined vocal features of an ethnically-uniform selection
of bilingual speakers (e.g. Altenberg and Ferrand,
2006; Jarvinen et al., 2013; Ng et al., 2010; Ng et al.,
2012; Todaka, 1993; Yamazawa and Hollien, 1992).
Parallel to the monolingual voice studies, however,
these bilingual studies failed to reach a consensus in
their findings of language difference. Whereas some
studies show promising results of the effect of language
on vocal parameters, especially F0 (Jarvinen et al.,
2013; Keating and Kuo, 2012; Ng et al., 2010; Ng
et al., 2012; Todaka, 1993; Yamazawa and Hollien,
1992), others failed to show a difference between
spoken languages in bilinguals (Altenberg and
Ferrand, 2006). Challenges included methodological
limitations (e.g. restricted selection of vocal measures;
order effects), limited speech sample selection, and
lack of consideration for second language factors.
With respect to factors that relate to the language
effect on bilingual voice, two major underlying
aspects in the use of different languages should be con-
sidered: the intrinsic linguistic features of each
language, and the contingencies of foreign language
learning. Both elements can affect the way the voice
is used in different languages, and thus influence the
measured outcomes of various vocal characteristics
in bilingual speakers.
Regarding the linguistic characteristics of language,
vocal aspects of speech can be influenced by com-
ponents such as tonal features (Altenberg and
Ferrand, 2006; Keating and Kuo, 2012; Ng et al.,
2010; Ng et al., 2012) or tense consonants (Jun,
1993). Tone languages such as Mandarin and
Cantonese include lexical tones that represent different
F0 features, which can relate to a larger range in F0
and, possibly, a higher average F0 during connected
speech. For example, Adrianopoulos et al. (2001)
reported in their study that Mandarin speakers
showed higher F0 and F0 ranges compared with
Hindi and American-English speakers when produ-
cing English vowels. Similar findings were found in
another study comparing Mandarin monolinguals
with English monolinguals (Keating and Kuo, 2012).
Additionally, tense consonants in Korean are under-
stood to function like lexical tones, with different F0
characteristics (Jun, 1993). Therefore, the influence
of consonants on F0 in Korean can be considered to

be tonal, as certain types of consonants predict a tonal
pattern. For instance, Korean introduces intrinsic
pitch features of tense and tone with the production
of certain consonants. The effect of tense consonants
on F0 in Korean is shown to spread across multiple
syllables, which can be represented phonologically
with tonal characteristics (Jun, 1993). Specifically, voi-
celess consonants and tense consonants correlate with
high tonal features, which can affect the overall F0
contours. Thus, different vocal patterns could possibly
be the result of the linguistic attributes of a language
itself.
As for the effect of second language use, degree of
language proficiency may have an impact on the oral
production of bilingual speakers (Jarvinen et al.,
2013; Ohala, 1984). A number of factors are claimed
to impact one’s second language proficiency such as
age of arrival (AoA) in a non-native-speaking
country, length of residence, age of second language
learning, amount of native language use, motivation,
and others (Asher and Garcia, 1969; Bialystok, 1997;
Flege and Fletcher, 1992; Flege et al., 1995;
MacKay et al., 2000; Oyama, 1976; Thompson, 1991).
Many of these variables are highly correlated to
degree of perceived foreign accent in the second
language (Asher and Garcia, 1969; Flege and
Fletcher, 1992), which may involve various aspects
of prosody, fluency, and articulation. Accent, slower
speech rate, increased F0, and low intensity in the
foreign language (Munro and Derwing, 2001) may
evoke different judgments of bilingual speakers. Slow
rate may evoke impressions of uncertainty and while
high f0 may arise from psychological anxiety, contri-
buting to tension in the laryngeal structures
(Jarvinen et al., 2013; Ohala, 1984). Soft speech can
signal passivity, insecurity, or compliance. For more
competent individuals, however, dual language lear-
ners may modify their speech characteristics to accom-
modate the second language, as a result of adaptation
and second language-related experience (Jarvinen
et al., 2013). These findings bring more attention to
vocal attributes in the bilingual population, which
may be altered according to language selection and
degree of second language competency.
The present investigation further highlights a notice-
able limitation in previous studies, variability in the
type of speech tasks, and attempts to address the
issue. The influence of task type is not a novel
concept in the literature on voice studies (Reich
et al., 1989; Zraick et al., 2005). Selection of different
speech materials can exert noticeable effects on vocal
features within individuals: specifically, many studies
have illustrated that an individual’s performance
level can vary, depending on whether a task is struc-
tured or unstructured, with the former producing
more advantageous results in performance (Baker
Lee and Sidtis The bilingual voice
et al., 2008; Hunter, 2009). Significantly different out-
comes in acoustic vocal measures were reported to be
highly task-oriented (Baker et al., 2008; Hunter,
2009; Keating and Kuo, 2012; Mang, 2001; Zraick
et al., 2005). Specifically, average F0 was significantly
higher in reading tasks than in spontaneous speech
samples across various populations of children,
young adults, older men, and adult trained singers
(Hudson and Holbrook, 1982; Mysak, 1959; Ramig
and Ringel, 1983; Sorenson, 1989). Also, it has been
observed that spontaneous speech constitutes a more
natural form of production in habitual voice than
other tasks, leading to difference in vocal character-
istics (Kreiman and Sidtis, 2011). Thus the selection
of task has been known to affect a speaker’s speech
performance in many aspects. However, a majority
of the bilingual voice studies restricted the conditions
of participants’ performance to a single task, and in
most cases, spontaneous speech was not included in
the design. The failure to obtain naturalistic speech
data when exploring aspects of vocal production
should not be disregarded. In order to capture the
voice profile of individuals in the most naturalistic
way, performance measures should be considered in
the broader context of connected speech.
This study aims to resolve the aforementioned issues
by investigating whether selection of language will
manifest different outcomes of vocal characteristics
across speech tasks in bilingual speakers. We
compare four vocal parameters in female Korean-
English and Mandarin-English bilingual individuals,
who are proficient in both languages, across different
speech tasks. Second language proficiency was evalu-
ated through personal interview by a certified ESL
teacher (BL), a language questionnaire, and a foreign
accent rating task. We selected two different groups
of bilingual speakers in order to study the effect of
using two distinct languages within an individual.
The measures under investigation for bilingual partici-
pants were (1) mean F0, (2) F0 variability, (3) intensity
(loudness), and (4) rate (syllables per second).
Methods
Participants
A total of 22 female bilingual speakers were recruited
to participate in the present study, consisting of 11
Korean-English speakers and 11 Mandarin-English
speakers. The Korean-English bilinguals (KE group)
had a mean age of 25.55 years (range = 21–27
years), and the Mandarin-English bilinguals (ME
group) had a mean age of 23.36 years (range = 22–
25 years). Participants’ education levels ranged from
baccalaureate to post-graduate. All of the selected par-
ticipants met the following selection criteria for bilin-
gualism: (a) their native language was either Korean
or Mandarin; (b) their parents were non-native
Speech, Language and Hearing 2017 3
Lee and Sidtis The bilingual voice
speakers of English; (c) they attended or were attend-
ing a school in the United States; (d) they were
fluent or proficient in English determined by an
initial screening interview by the first author (BL, a
fluent Korean-English bilingual, has teacher certifica-
tion in ‘teaching English as a second language’).
Second language proficiency is defined in many
terms and used differently by various researchers
(for a review, see McNamara, 1996). Among the
different components of competency in second
language, oral proficiency was considered the crucial
determining factor that reflects different patterns of
vocal features in speech production. Thus, partici-
pants’ English oral competency was determined by
two conditions: an oral screening interview, and
foreign accent rating in English. For the oral screening
interview, participants were responded to questions in
English such as ‘How are you feeling today?’ or ‘How
did you find out about the study?’ Potential partici-
pants with nonfluent, inaccurately expressed, or
ungrammatical responses were excluded from the
study.
The KE group reported speaking the Seoul dialect of
Korean, and Modern Standard Chinese (Mandarin,
Putonghua) was spoken by the ME group. All partici-
pants attested that their voice sounded normal and that
their health status was good or excellent on the day of
testing. Written consent was obtained from all partici-
pants using the Human Subjects Consent Forms, as
mandated and approved by the Internal Review
Board of the University.
Recording procedure and tasks
Participants’ speech was recorded in a sound-attenu-
ated booth in the Speech and Hearing Clinic at
New York University. At the end of each speech
recording session, a written questionnaire was admi-
nistered to participants to obtain information on
their language and educational background. A
description of the demographic information of the
participants is displayed in Table 1.
All participants performed three different speech
tasks: a reading passage, a monologue, and a picture
description, each in their two spoken languages. All
instructions were provided in written form. The
research design consisted of two experimental blocks,
wherein the use of only one language was permitted
per block. To control for order effects, the order of
language was counterbalanced across participants
and within each language, and the order of tasks was
randomized. In all, there were six possible presentation
orders, combining the order of the language and task.
Reading task
A reading passage was provided to the participants in
English and in their native language. Each participant
4 Speech, Language and Hearing 2017
was asked to read ‘The Rainbow Passage’ (Fairbanks,
1960) at a normal speech rate, using a comfortable
speaking voice. For Mandarin-English speakers, a
translated version of ‘The Rainbow Passage’ was pro-
vided in Mandarin. For Korean-English speakers, the
same reading passage was translated in Korean. All
participants were allowed to practice reading the
passage a number of times to familiarize themselves
with the task, before the speech samples were
recorded.
Picture description
The ‘Cookie Theft’ picture of the Boston Diagnostic
Aphasia Examination (Goodglass and Kaplan, 1972)
was used to elicit a narrative discourse in the form of
a picture description. Written instructions were pro-
vided in both the native language and English. Each
speaker was allowed sufficient time to examine the
picture before elicitation. The same picture was used
in both spoken languages for each bilingual.
Monologue
To elicit participants’ spontaneous connected speech,
all speakers were instructed to talk briefly about a
selected topic, either an ‘introduction to their
family’, or their ‘life experience in United States’.
The speech sample was obtained for a period of
approximately one to two minutes. The same topic
was used for speech samples in both the native
language and English for each participant.
Recordings
All speech samples in the experiment were recorded on
a digital audiotape (DAT) recorder (Tascam DR-40)
and sampled in a sound-attenuated booth using a
high quality head-mounted microphone (Sennheiser).
The recordings were digitized at a sampling rate of
44. 1 kHz and 16 bits/sample quantization, and
stored in a laptop for subsequent analyses. The micro-
phone-to-mouth distance for each talker was main-
tained to be approximately 3 – 4 cm at all times, and
slightly to the side of the speaker’s mouth.
Accent rating in English
In addition to the selection of participants based on
screening interviews, the degree of foreign accent in
English was assessed for each speaker. The purpose
of this task was to further determine oral language
proficiency by evaluating the degree of detectable
non-native accent in second language production. A
listening test was devised in order to obtain measures
of degree of accent in the participating speakers.
The rating task was performed by 20 native speakers
of English (three males, 17 females). Their ages ranged
from 21 to 51 years (a mean of 31.9 years, SD = 4.35).
All listeners were born and educated in the United
States and ranged in education from 12 to 18 years

Table 1 Demographic information on participants in Korean-English (KE) group and Mandarin-English (ME) group; mean values
and standard deviation (in parentheses) are provided
General English
Total education learning
Group (n) Age (years) (years)
KE 11 25.5 (1.75) 17.18 (1.25) 13.55 (2.83)
ME 11 23.4 (0.81) 17.27 (0.47) 13.32 (3.19)
(SD = 1.89). Participants with or without a back-
ground in communicative sciences were included.
Participants were excluded if they had prior language
knowledge on Mandarin or Korean. Each listener par-
ticipated in the experiment privately in a quiet room
using a software program EXPERIGEN (Becker and
Levine, 2014) that enabled implementation of the
study and data collection via computer interface. All
listeners were informed that a series of speech
samples would be played to them one at a time and
their task was to assess how strong they perceived
the foreign accent to be. Based on a 9-point Likert
scale (1 – very strong foreign accent, 9 – no foreign
accent), a total of 132 speech stimuli was assessed by
listeners. After hearing an utterance, a visual scale
from 1 to 9 appeared on the computer screen and
each participant clicked on a number that matched
to their decision of degree of accent. Before the
testing started, two practice trials were provided with
speech samples (which were excluded from the exper-
iment) produced by native speakers of English.
The stimuli in the experiment were 6–10 second
speech recordings excerpted from the original exper-
iment, all representing the Korean-English and
Mandarin-English speakers’ speech in English. Three
utterances were selected each from the reading task
and spontaneous speech (a total of six utterances per
speaker), respectively. This was intended to examine
whether there would be a difference between the
speech tasks in perception of accent and to provide
additional information on the bilingual speakers. It
has been reported that reading is rated as more strongly
accented than spontaneous speech (Oyama 1976;
Thompson, 1991) and we intended to examine whether
the task effect is maintained in our accent ratings. The
order of presentation of the trials was randomized for
each listener, and the stimuli could be replayed if needed.
Data analysis
Acoustic measures
Acoustic vocal measures, including average F0, F0
variability (coefficient of variation; standard deviation
divided by mean), average intensity (dB SPL), and
speech rate (syllables per sec) were obtained from the
speech sample recordings. Regarding speech rate,
since factors such as dysfluencies and pause-fillers
Lee and Sidtis The bilingual voice
Self-rated Self-rated
English oral overall English
Age of arrival Years of fluency (1 – very proficiency
(AoA) in the residence in poor, 5 – near- (1 – very poor,
U.S. the U.S. native) 5 – near-native)
12.91 (7.25) 10.59 (4.33) 4.18 (0.87) 4.36 (0.67)
22.18 (0.98) 1.22 (0.41) 3.55 (0.69) 3.73 (0.65)
are frequently detected in natural speech
(Cucchiarini et al., 2002), values were obtained only
for the reading task. Measures were acoustically ana-
lyzed from the recorded speech samples utilizing
Praat (Boersma, 2001), a speech analysis software
program, using a pitch tracking system with a standard
range setting of 75– 500 Hz. For acoustic data analy-
sis, participants’ responses were segmented into mul-
tiple phrasal or sentential utterances. Silence periods
between phrases and sentences were edited if they
were beyond one second. Values of the measures for
all segmented portions of connected speech were
obtained and averaged. All measures were manually
checked and corrected for artifacts (e.g. spurious F0
values outside the range of normal pitch).
Statistical analyses
For the accent rating task, a two-way mixed analysis of
variance (ANOVA), using SPSS software program for
statistical analyses, was conducted to compare the
degree of foreign accent between KE and ME group
in the reading task and monologue. Additionally, for
group comparison, independent t-tests and chi-
square tests were conducted on the demographic infor-
mation of the KE and ME groups. Regarding the
acoustic measures, a series of 3-way mixed ANOVA
and 2-way mixed ANOVA tests were conducted on
the data. The measures of F0, F0 variability, and
intensity were analyzed using three-way mixed
ANOVA with Language (native language vs. English)
and Task (reading, picture description, monologue)
as within-group variables, and Group (KE group vs.
ME group) as a between-group variable. Speech rate
was analyzed using a two-way mixed ANOVA with
Language (native language vs. English) and Group
(KE group vs. ME group) as factors. We considered
an alpha level of 0.05 to be statistically significant in
all following analyses. The Bonferroni’s test was used
to carry out post hoc pairwise comparisons, when
the ANOVA results indicated significant effects
within or between variables.
Results
Demographic comparisons
Comparisons were made between the KE and ME
groups based on the data obtained in the written
Speech, Language and Hearing 2017 5
Lee and Sidtis The bilingual voice
questionnaire (see Table 1). The following factors of
interest were analyzed: age, years of general education,
years of English education, AoA in the States, years of
residence in the States, self-rated English oral profi-
ciency (how fluent one speaks in English), self-rated
overall English competency (overall command of
English). Self-ratings on the participants’ English
competency and oral proficiency were made on a
scale from 1 (very poor) to 5 (near-native).
Results revealed that KE group and ME group did
not differ in number of years of general education
and years of English learning. Also, for self-rated
English oral proficiency, independent t-tests failed to
reveal significant differences between the KE and
ME groups (P = 0.149). For the scores on self-rated
English overall proficiency, however, the results
yielded a significant difference between the KE and
ME groups (P = 0.04), as the KE group perceived
their overall proficiency in English to be better than
the ME group. The two groups showed significant
differences in three additional categories: age (P <
0.001), AoA in the US (P < 0.005), and years of resi-
dence in the US (P < 0.005). Regarding age, the
mean age of the ME group was slightly lower than
the mean age of the KE group (gap of 2.1 years),
but the participants’ ages ranged between 21 and 27
years, which are not expected to display distinct
vocal differences due to physiological changes based
on aging effect (Boulet and Oddens, 1988; Sataloff
and Linvelle, 2006). The average AoA of KE group
was lower than ME group, indicating that the
Korean-English bilinguals were exposed to a richer
English environment, arriving in the U.S. at a mean
age of 12.91 years, whereas the participants in ME
group arrived in the U.S. at a mean age of 22.18
years. A corresponding difference occurred in the
years of residence in the U.S. also. Specifically, KE
group had an average of 10.59 years (SD = 7.25) for
Figure 1 Ratings of the degree of accent in English between
KE and ME group across two speech tasks are shown in line
graphs, respectively. Results show significant difference
between KE and ME group across both tasks, and between
reading and spontaneous speech for KE group only (P < 0.05).
6 Speech, Language and Hearing 2017
the years of residence, which was significantly higher
than the mean of ME group of 1.22 years (SD = 0.41).
Overall, group comparison results reveal that both
groups were fairly competent in oral production of
their second language, considering English education
years and self-ratings of English oral fluency.
However, differences were noticed between the two
bilingual groups such as AoA, years of residence in
the U.S., and self-ratings of English proficiency.
Along with the results for the accent ratings, this
issue will be further addressed in the discussion
section.
Accent rating in English
The foreign accentedness in the English speech
samples of KE and ME group were analyzed using a
two-way mixed ANOVA with Task (reading vs. spon-
taneous) and Group (KE vs. ME group) as factors.
The mixed ANOVA results yielded significant main
effects of Task (F (2,20) = 7.297, P < 0.005) and
Group (F (1,20) = 1468.518, P < 0.005), but no sig-
nificant interaction effect of Task × Group
(F (2,20) = 0.165, P = 0.847). Fig. 1 depicts the
average rating values across tasks and groups. As
shown, KE group had a significantly higher rating
(less accented speech) compared to the ME group in
both tasks. This indicates that, despite both groups
showing high level of English on other measures, the
accent rating task provided further information on
second language proficiency in the oral production,
with the KE group perceived as having less foreign
accent than the ME group. For the KE group, a sig-
nificant task difference between accent ratings was
also found; accent was rated on an average of 4.37
(SD = 0.27) for reading and 3.93 (SD = 0.40) for
spontaneous speech (t(20) = 2.007, P = 0.06), indicat-
ing that the speech samples from the reading task was
perceived as being less accented than spontaneous
speech. For the ME group, however, the average
rating of their English was 6.78 (SD = 0.15) in
reading, whereas they were rated at a mean of 6.53
(SD = 0.54) for spontaneous speech, with no signifi-
cant difference (t(20) = 2.246, P = 0.54).
The two groups differed in their native languages,
and therefore the effects of language characteristics
on perceived pronunciation cannot be ruled out.
However, the rating differences were likely influenced
at least in part by demographic factors. It can be
assumed from the demographic data, for whom some
pertinent characteristics are equated, that AOA and
that years of residence in the U.S. may have influenced
the results of the rating protocol (Flege et al., 1995).
Mean F0
A three-way mixed ANOVA revealed a significant
Language effect, F(1, 20) = 14.79, P = 0.001 on F0

measures, indicating an overall significantly higher
average F0 in the native language compared to
English (Korean, M = 213.61, SD = 3.21; Mandarin,
M = 205.38, SD = 2.91) in both KE and ME groups.
A significant main effect of Task was also found (F
(2, 20) = 37.4, P < 0.0005), with the reading task
having the highest mean F0 values, followed by the
picture description and the monologue task. Post hoc
pairwise comparisons revealed that there was a signifi-
cant difference between reading and picture descrip-
tion (P < 0.0005), and reading and monologue (P <
0.0005), but picture description and monologue did
not show significant differences on F0 values. There
was no main effect of Group or interaction between
Group and Language, indicating that there was no
overall group difference on the average F0 values (F
(1,20) = 0.67, P = 0.421) nor was there a group differ-
ence between the F0s of native language and English
(F(1,20) = 1.792, P = 0.196). However, the results
yielded significant interaction effects of Language ×
Task, F(2,20) = 5.357, P = 0.009), and Group × Task,
F(2,20) = 8.624, P = 0.001, reflecting that, depending
on the given task, there was a difference between
native language and English, and the groups differed
in their performance between the two spoken
languages; post hoc tests showed that the KE group
consistently exhibited a significantly higher mean F0
in Korean compared with English in all three tasks;
for the ME group, however, the average F0 was signifi-
cantly higher in Mandarin than in English only in the
reading task (P = 0.014), indicating that differences in
F0 between speaking Mandarin and English depended
on the type of speech task. The effects of F0 values
associated with spoken languages and tasks in the
KE and ME groups are depicted in Fig. 2.
F0 variability
To measure F0 variability, coefficient of variation
(standard deviation divided by mean) was used to esti-
mate the relative magnitude of the variance from the
mean F0. Results indicated a significant main effect
of Language (F(1,20) = 6.073, P = 0.023). Post hoc
pairwise comparisons showed that the KE group
exhibited a significantly greater pitch variability in
Korean compared with English (P = 0.01), as depicted
in Fig. 3. For the ME group, F0 variability was similar
between Mandarin and English for all three tasks.
There were no significant effects of Group or Task,
or any significant interaction effects on F0 variability,
indicating that task selection did not affect the varia-
bility in F0 for both groups.
Intensity
ANOVA results revealed a significant main effect of
Group (F(1,20) = 11.84, P < 0.003) on intensity.
Fig. 4 illustrates the average intensity (dB) values for
Lee and Sidtis The bilingual voice
the two spoken languages across all tasks for both
groups. However, neither Language nor Task, or
their interaction, had any impact on intensity. The
ME group showed a higher average intensity com-
pared to the KE group with a mean difference value
of 4.99 dB, indicating that the ME group were
measured as significantly louder, but the intensity
between native language and English was comparable
across tasks and languages.
Speech rate
The 2 × 2 mixed ANOVA results yielded significant
main effects of Language (F(1,20) = 48.678, P <
0.0005) and Group (F (1,20) = 12.283, P < 0.0005),
and a significant interaction effect of Language ×
Group (F (1,20) = 12.283, P = 0.002). Fig. 5 shows
the average rate (syllables per second) across native
language and English. Participants spoke significantly
faster in their native language compared to English.
For KE group, Korean was spoken at a rate of 5.18
syllables per second (SD = 0.45), whereas English
was spoken at a mean of 4.03 syllables per second
(SD = 0.54). For ME group, average rate for
Mandarin was 4.01 syllables per second (SD = 0.27)
and 3.63 syllables per second (SD = 0.40) for
English. Results showed that speech rates between
KE and ME group also differed significantly. The
rate difference between the two languages was
greater in the KE group than in the ME group.
When rates for English were compared using t-tests,
however, there was only a marginally significant differ-
ence (t(20) = 2.007, P = 0.06), indicating that the
speech rate for English was comparable between
groups.
Discussion
This study examined the voice profiles of bilingual
speakers when speaking two different languages
across speech tasks in order to address the possible
difference in vocal parameters related to language
selection and type of speech task. F0, F0 variability,
intensity, and rate, parameters of importance in
judging a voice pattern, were the measures selected
for this study. Findings indicated that there was a sig-
nificant effect of language and task on these selected
vocal characteristics in bilingual speakers. With
respect to language effect, the bilingual speakers
showed significant differences in the average F0, F0
variability, and speech rate in their two languages.
Furthermore, there was an interaction effect between
language selection and task on F0, suggesting that
the difference in F0 between languages was based on
the speech task that the bilinguals performed.
However, intensity did not reveal any language or
task effect, but showed the ME group speaking
overall significantly louder than the KE group.
Speech, Language and Hearing 2017 7
Lee and Sidtis The bilingual voice

Figure 2 Average F0 (Hz) and F0 standard deviations (error bars) between two spoken languages across three speech tasks are
shown in bar graphs for KE (left) and ME (right) group, respectively. Statistical significance is indicated by an asterisk (P < 0.05).

Speech rate differed between two spoken languages
within both groups, but also between the groups.
Overall, the findings provide further support that the
bilingual voice features consistently different vocal
patterns, contingent not only on the language (native
versus acquired) that is produced, but also on the
type of speech task. We turn to provide some expla-
nations of the outcomes and for the resulting profile
of a bilingual voice.
Concerning F0 measures, a significantly higher F0
in either Korean or Mandarin than in English was
evident – concurring with findings reported in some
previous studies (Keating and Kuo, 2012; Mang,
2001). This contrasts with other bilingual studies that
found a higher F0 in English than in the native
languages (Jarvinen et al., 2013; Ng et al., 2010; Ng
et al., 2012), suggesting that using a foreign language
can lead to possible higher F0 due to emotional
factors (Jarvinen et al., 2013). It is speculated that all
participants in the study were overall fluent and suffi-
ciently competent in oral English so that the effect of
speaking a foreign language did not strongly affect a
change in the F0 (i.e. increased F0). In that regard,
the average F0 values in English between the two
groups were not significantly different, suggesting
that the difference in F0s based on language selection
cannot be fully explained by simply the influence of
using a second language.

Figure 3 Utterance-level mean F0 variability (coefficient of variation) between two spoken languages across speech tasks is
shown in bar graphs for KE (left) and ME (right) group, respectively. Statistical significance is indicated by an asterisk (P < 0.05).

8 Speech, Language and Hearing 2017

 Lee and Sidtis The bilingual voice

Figure 4 Utterance-level mean intensity (dB) between native language and English across three speech tasks is shown in bar
graphs for KE (left) and ME (right) group, respectively. Statistical significance is indicated by an asterisk (P < 0.05).

Rather, other factors should be of consideration. If a
bilingual speaker is orally proficient in two languages,
it may be that the intrinsic acoustic features of each
language contribute to the difference in the acoustic
vocal parameters. Specifically, acoustic phonetic
characteristics of each language may have influenced
vocal changes in overall speech. As mentioned, a
higher F0 for tonal languages was demonstrated, com-
pared with non-tonal languages, not only during iso-
lated vowel productions (Eady, 1982; Mang, 2001)
but also in connected speech (Keating and Kuo,
2012; Ng et al., 2012). The current findings support
the notion of a normative higher F0 in tonal languages
(Mandarin) than in non-tonal ones (English). In the
case of the Korean language, also, the mean F0 in
Korean was significantly higher than in English.
Although there are no reported cross-linguistic
studies that compare the acoustic pitch features of
Korean and English in connected speech, we consider
the consonant-tone interaction and tense consonants
in Korean to be a possible cause of the similar pho-
netic effects of higher F0, in a manner parallel to
that of tonal languages. Considering that tense conso-
nants in Korean but not in English bring forth a higher
F0, the difference in F0 between Korean and English
could be attributed to the intrinsic F0 features of
tense and tonal effect of Korean consonants. The
Korean-English bilingual speakers may adequately

Figure 5 Utterance-level mean speech rate (syllables per second) and standard deviations (error bars) of speaking rate
between native language and English in reading task are shown in graphs for KE (left) and ME (right) group, respectively.
Statistical significance is indicated by an asterisk (P < 0.05)*.

Speech, Language and Hearing 2017 9

Lee and Sidtis The bilingual voice
produce the tense consonants and tonal phonetic
effects of consonants in Korean during connected
speech, resulting in overall higher F0. We speculate
that the vocal differences between the two spoken
languages in the KE group may derive from possible
linguistic factors such as this difference in phonetic
repertoire of consonants. Therefore, both Korean
and Mandarin resulted in overall higher F0 than
English. Further studies that explore the relationship
between tone and tense in languages and overall
pitch change in the natural speech of bilingual speak-
ers should be considered.
Interestingly, it is noted that whereas the KE group
showed significant differences between their two
spoken languages across all speech tasks, the ME
group demonstrated significant difference only
during the reading task. There can be two explanations
for the different F0 outcomes across groups. First,
reading is a structured task that proven to be a
robust measure for eliciting comparatively higher F0
values (Hudson and Holbrook, 1982; Mysak, 1959;
Ramig and Ringel, 1983; Sorenson, 1989). It may be
that the nature of performing a reading task can
underscore the acoustic attributes of each language
since the speakers may hyper-articulate with better
control of speech, resulting in a stark F0 contrast
between the two spoken languages. Another reason
may lie in the difference in oral English proficiency
observed in the accent ratings. Although overall
English skills were competent among the two groups,
a stronger foreign accent in English for the ME
group may have led to comparable F0s between
Mandarin and English on their spontaneous speech
provide partial explanation to the F0 difference in
two languages across tasks. Various aspects of oral
proficiency in a language may certainly have an
impact on the acoustic patterns in the bilingual
voice. The effect of second language oral proficiency
on voice in many of its perspectives remains to be
explored further.
With regard to speech rate, there was a significant
difference between the two spoken languages for
both groups. This can be attributed to similar issues
dealt in the results of either the intrinsic linguistic fea-
tures of each language or F0 relating to second
language oral proficiency. Regarding the first reason,
a recent study conducted by Pellegrino and her col-
leagues (2011) shed light on how the speed of spoken
production can be related to syllabic rate and infor-
mation density. According to their study, there exists
a negative correlation between speaking rate based
on syllabic structure and syllabic information density.
Languages that have a high speaking rate have a ten-
dency to pack less semantic information into each syl-
lable (decreased information density) and vice versa.
For instance, Spanish has a high speech rate with
10 Speech, Language and Hearing 2017
low density syllables, whereas Mandarin is considered
a language with slower syllabic rate with high infor-
mation density (due to tonal features packed in sylla-
bles). The present study is in agreement with
Pellegrino et al. (2011) in that Korean exhibits
higher syllabic speaking rates compared with English
and Mandarin, which are characterized by slow sylla-
bic rate with high density syllables. Although the infor-
mation rate theory does not provide a comprehensive
explanation of the overall acoustic differences, it
suggests that language-specific syllable density may
influence measures of speaking rate. Considering the
speech rate findings to be related to second language
use, it is mentioned that overall second language pro-
ficiency can influence the rate in natural speech
(Munro and Derwing, 2001). That is, a slower rate in
the second language, compared to the native tongue,
is the result of the higher demands of the cognitive
task and a lack of competency in oral control. We con-
firmed in this study that both groups demonstrated a
slower rate in English, which may be attributed to
the fact that all of the speakers are non-native speakers
of English. The F0 measure may not have been
affected by the second language proficiency, but it
may have for the speech rate. Further investigation is
warranted to determine how linguistic factors influ-
ence speaking rates across languages.
Another possible explanation for the different
voice profiles in bilinguals results from a consider-
ation of socio-cultural or pragmatic factors that can
influence vocal features within a language (van
Bezooijen, 1995). Changing vocal quality within
individuals can occur with various pragmatic func-
tions (Kreiman and Sidtis, 2011; Zraick et al.,
2005). Loveday (1981), for instance, explored F0
differences in Japanese and English speakers in
both males and females, and found that Japanese
females adopted a high F0 that is very distinct
from both English speakers and Japanese males.
This was attributed to socio-cultural expectations
and the pragmatic purpose of expressing politeness
or femininity. Studies on the phonetic profile of
politeness revealed that Korean native speakers
also show a marked distinction in average F0
between different levels of polite speech, but in this
case greater politeness was correlated with lower
average F0 (Brown et al., 2014; Shin, 2005; Winter
and Grawunder, 2012). Nonetheless, the fact that
Korean female speakers share with Japanese
females the cultural values of exhibiting expected
politeness and femininity within society (Han,
1992) indicates that socio-cultural factors may influ-
ence changes in vocal quality. The potential corre-
lation between cultural characteristics in certain
language groups and acoustic vocal features of
speech seems to merit further investigation.

In respect to task effect on vocal features, early
studies on voice have demonstrated that monolingual
speakers have a higher F0 associated with a structured
task compared with an unstructured one (Hollien
et al., 1997; Hollien and Jackson, 1973). Our findings
show that this effect of task applies consistently to
bilinguals. This observation was also reflected in the
accent ratings, where reading was rated as significantly
less accented than spontaneous speech. Task effect
may be evident in both perception and production of
speech. The inconsistent findings from the previous lit-
erature may have resulted from the restricted or selec-
tive methodological approach of using a specific task
without comparison to other tasks.
The present study is distinct from previous works by
its inclusion of a number of vocal parameters and the
attention to speech task type in establishing a more
refined bilingual voice profile. However, caution
should be taken not to generalize the findings to the
entire bilingual population, since only a limited
number of females were included in the study; there
may well be differences between genders in use of
voice on tasks. Furthermore, as mentioned above,
although language competency in the two spoken
languages was overall well-balanced, subtle differences
in oral English pronunciation skills occurred between
groups. Based on the language profiles of the ME
group, this is somewhat predictable as there is a posi-
tive correlation between detectable foreign accents
and the age of learning in a second language (Flege
and Fletcher, 1992; Thompson, 1991). Accented pro-
duction in a language can induce differences in
phoneme sets and intonation patterns, and create
overall changes in pronunciation patterns (Arslan
and Hansen, 1996). It is not known to what extent
degree of foreign accent influences vocal character-
istics in bilinguals and this merits exploration in
future studies. In order to do so, comparison of
groups of monolinguals, and bilinguals with same
and different native language, should be explored in
details. Despite caveats, the findings here provide
support for the growing body of literature showing
that the use of different languages clearly manifests
different outcomes of vocal characteristics, even
within bilingual speakers.
Conclusion
The present study attempts to explore vocal differences
in bilingual speakers that occur, depending on their
choice of language and type of speech task. Findings
revealed that bilingual speakers produced notably
different voice patterns in their two spoken languages
across different speech tasks, when vocal measures
were explored. The difference implies that language
alone is an acquired factor that contributes to the
manifestation of within- and between-speaker
Lee and Sidtis The bilingual voice
variability of vocal attributes. In addition, results indi-
cate that selection of speech sample can be a crucial
factor in differentiating acoustic profiles, supporting
previous literature that task effects emerge in acoustic
measures within individuals (Zraick et al., 2005).
Future studies will do well to further investigate
vocal features of bilingual speech relating to factors
such as gender, age, and language dominance, for a
better understanding of bilingualism and its effect on
natural speech.
Disclaimer statements
Contributors None.
Funding None.
Conflicts of interest None.
Ethical approval None.
ORCID
Binna Lee http://orcid.org/0000-0001-5320-9703
References
Altenberg, E.P., Ferrand, C.T. 2006. Fundamental frequency in
monolingual English, bilingual English/Russian, and bilingual
English/Cantonese young adult women. Journal of Voice, 20:
89–96.
Amino, K., Osanai, T. 2015. Cross-language differences of articula-
tion rate and its transfer into Japanese as a second language.
Forensic Science International, 249: 116–122.
Andrianopoulos, M.V., Darrow, N.K., Chen, J. 2001. Multimodal
standardization of voice among four multicultural populations:
fundamental frequency and spectral characteristics. Journal of
Voice, 15: 259–272.
Arslan, L., Hansen, J. 1996. A study of temporal features and fre-
quency characteristics in American English foreign accent.
The Journal of the Acoustical Society of America, 102(1): 28–40.
Asher, J.J., García, R. 1969. The optimal age to learn a foreign
language. The Modern Language Journal, 53(5): 334–341.
Awan, S.N., Mueller, P.B. 1996. Speaking fundamental frequency
characteristics of white, American, and Hispanic kindergarten-
ers. Journal of Speech, Hearing, and Research, 39: 573–577.
Baken, R., Orlikoff, R. 2000. Clinical measurement of speech and
voice. San Diego, CA: Singular Thomson Learning.
Baker, S., Weinrich, B., Bevington, M., Schroth, K., Schroeder, E.
2008. The effect of task type on fundamental frequency in chil-
dren. International Journal of Pediatric Otorhinolaryngology, 72
(6): 885–889.
Becker, M., Jonathan, L. 2014. Experigen – an online experiment
platform. Available at http://becker.phonologist.org/
experigen.
Bialystok, E. 1997. The structure of age: in search of barriers to
second language acquisition. Second Language Research, 13
(2): 116–137.
Boersma, P. 2001. Praat, a system for doing phonetics by computer.
Glot International, 5: 341–345.
Boulet, M.J., Oddens, B.J. 1988. Female voice changes around and
after the menopause – an initial investigation. Maturitas, 23(1):
15–21.
Brown, L., Winter, B., Idemaru, K., Grawunder, S. 2014. Phonetics
and politeness: perceiving Korean honorific and non-honorific
speech through phonetic cues. Journal of Pragmatics, 66: 45–60.
Cucchiarini, C., Strik, H., Boves, L. 2002. Quantitative assessment
of second language learners’ fluency: comparisons between
read and spontaneous speech. The Journal of the Acoustic
Society in America, 111(6): 2862–2873.
Eady, S. 1982. Differences in the F0 patterns of speech: tone
language versus stress language. Language and Speech, 25:
29–42.
Fairbanks, G. 1960. Voice and articulation drillbook. New York:
Harper and Row.
Speech, Language and Hearing 2017 11
Lee and Sidtis The bilingual voice
Flege, J., Fletcher, K. 1992. Talker and listener effects on the percep-
tion of degree of foreign accent. The Journal of the Acoustic
Society in America, 91: 370–389.
Flege, J., Munro, M., Mackay, I. 1995. Factors affecting strength of
perceived foreign accent in a second language. The Journal of
the Acoustic Society in America, 97: 3125–3134.
Gelfer, M., Denor, S. 2014. Speaking fundamental frequency and
individual variability in Caucasian and African-American
school-aged children. American Journal of Speech-Language
Pathology, 23(3): 395–406.
Goodglass, H., Kaplan, E. 1972. The assessment of aphasia and
related disorders. Lea and Febiger: Philadelphia.
Han, C. 1992. A comparative study of compliment responses:
Korean females in Korean interactions and in English inter-
actions. Working Papers in Educational Linguistics, 8: 17–31.
Hanley, T.D., Snidecor, J.C., Ringel, R.L. 1966. Some acoustic
differences among languages. Phonetica, 14: 97–107.
Hollien, H., Jackson, B. 1973. Normative data on the speaking fun-
damental frequency characteristics of young adult males.
Journal of Phonetics, 1: 117–20.
Hollien, H., Hollien, P.A., Jong, G. 1997. Effect of three parameters
on speaking fundamental frequency. The Journal of the
Acoustical Society of America, 102(5): 2984–2992.
Hudson, A.I., Holbrook, A. 1982. Fundamental frequency charac-
teristics of young black adults: spontaneous speaking and oral
reading. Journal of Speech and Hearing Research, 25: 25–28.
Hunter, E. 2009. A comparison of a child’s fundamental frequencies
in structured elicited vocalizations versus unstructured natural
vocalizations: a case study. International Journal of Pediatric
Otorhinoloaryngology, 73: 561–571.
Jarvinen, K., Laukkanen, A., Aaltonen, O. 2013. Speaking a foreign
language and its effect on F0. Logopedics Phoniatrics Vocology,
38: 47–51.
Jun, S. 1993. The phonetics and phonology of Korean prosody
(Doctoral dissertation, The Ohio State University).
Keating, P., Kuo, P. 2012. Comparison of speaking fundamental fre-
quency in English and Mandarin. Journal of Acoustic Society of
America, 132: 1050–1060.
Kreiman, J., Sidtis, D. 2011. Foundations of voice studies: an interdis-
ciplinary approach to voice production and perception. Boston:
Wiley-Blackwell.
Loveday, L. 1981. Pitch, politeness and sexual role: an exploratory
investigation into the pitch correlates of English and Japanese
politeness formulae. Language and Speech, 24: 71–89.
MacKay, I.R., Meador, D., Flege, J.E. 2000. The identification of
English consonants by native speakers of Italian. Phonetica,
58(1–2): 103–125.
Mang, E. 2001. A cross-language comparison of preschool chil-
dren’s vocal fundamental frequency in speech and song pro-
duction. Research Studies in Music Education, 16: 4–14.
Mayo, R., Manning, W.H. 1994. Vocal tract characteristics of
African-American and European-American adult males.
Texas Journal of Audiology and Speech-Language Pathology,
20: 33–36.
McNamara, T.F. 1996. Measuring second language performance.
London: Addison Wesley Longman.
Mennen, I., Schaeffler, F., Docherty, G. 2012. Cross-language differ-
ences in fundamental frequency range: a comparison of English
and German. The Journal of the Acoustic Society of America,
131: 2249–2260.
Morris, R.J. 1997. Speaking fundamental frequency characteristics
of 8- through 10-year-old white- and African-American boys.
Journal of Communication Disorders, 30: 101–116.
12 Speech, Language and Hearing 2017
Munro, M.J., Derwing, T.M. 2001. Modeling perceptions of accent-
edness and comprehensibility of L2 speech. Studies in Second
Language Acquisition, 23: 451–468.
Mysak, E.D. 1959. Pitch and duration characteristics of older males.
Journal of Speech and Hearing Research, 2: 46–54.
Ng, M., Chen, Y., Chan, Y.K. 2012. Differences in vocal character-
istics between Cantonese and English produced by Cantonese-
English bilingual speakers – a long-term spectral analysis.
Journal of Voice, 26: e171–e176.
Ng, M., Hsueh, G., Leung, C. 2010. Voice pitch characteristics of
Cantonese and English produced by Cantonese-English bilin-
gual children. International journal of speech-language pathol-
ogy, 12(3): 230–236.
Ohala, J. 1984. An ethological perspective on common cross-
language utilization of F0 of voice. Phonetica, 41: 1–16.
Oyama, S. 1976. A sensitive period for the acquisition of a nonnative
phonological system. Journal of Psycholinguistic Research, 5(3):
261–283.
Pellegrino, F., Coupe, C., Marsico, E. 2011. Across-language per-
spective on speech information rate. Language, 87(3): 539–558.
Ramig, L.A., Ringel, R.L. 1983. Effect of physiological aging on
selected acoustic characteristics of voice. Journal of Speech
and Hearing Research, 26(1): 22–30.
Reich, A.R., Mason, J.A., Frederickson, R.R., Schlauch, R.S. 1989.
Factors influencing fundamental frequency range estimates in
children. Journal of Speech and Hearing Disorders, 54(3): 429–438.
Roberts, P.M., Shenker, R.C. 2007. Assessment and treatment of
stuttering in bilingual speakers. In: Conture E, Curlee R.F.
(eds.) Stuttering and related disorders of fluency (3rd ed.) New
York: Thieme Medical Publishers, Inc., pp. 183–209.
Sapienza, C.M. 1997. Aerodynamic and acoustic characteristics of the
adult African American voice. Journal of Voice, 11: 410–416.
Sataloff, R.T., Linvelle, S.E. 2006. Vocal health and pedagogy:
advanced assessment and treatment. 2nd ed. San Diego: Plural.
Shin, S. 2005. Grammaticalization of politeness: a contrastive study of
German, English and Korean. Berkeley: Doctoral thesis,
University of California.
Sorenson, D.N. 1989. A fundamental frequency investigation of
children ages 6–10 years old. Journal of Communication
Disorders, 22(2): 115–123.
Thompson, I. 1991. Foreign accents revisited: the English pronuncia-
tion of Russian immigrants. Language Learning, 41: 177–204.
Todaka, Y. 1993. A cross-language study of voice quality: bilingual
Japanese and American speakers. Los Angeles: Doctoral disser-
tation, University of California.
van Bezooijen, R. 1995. Sociocultural aspects of pitch differences
between Japanese and Dutch women. Language and Speech,
38(3): 253–256.
Wheat, M.C., Hudson, A.I. 1988. Spontaneous speaking fundamen-
tal frequency of 6-year-old black children. Journal of Speech,
Language, and Hearing Research, 31(4): 723–725.
Winter, B., Grawunder, S. 2012. The phonetic profile of Korean
formal and informal speech registers. Journal of Phonetics,
40(6): 808–815.
Xue, S.A., Needley, R., Hagstrom, F., Hao, J. 2001. Speaking F0
characteristics of elderly Euro-American and African-
American speakers: building a clinical comparative platform.
Clinical Linguistic Phonetics, 15: 245–252.
Yamazawa, H., Hollien, H. 1992. Speaking fundamental frequency
patterns of Japanese women. Phonetica, 49: 128–140.
Zraick, R.I., Birdwell, K., Smith-Olinde, L. 2005. The effect of
speaking sample duration on determination of habitual pitch.
Journal of Voice, 19: 197–201.